Ohio's supercomputer: 154 trillion calculations per second
Better. Stronger. Faster.
These three adjectives once described television's fictional astronaut Steve Austin after $6 million in bionic implants transformed him into a superhuman. Today, those same three words could illustrate the advanced--and very real--computational abilities of the Ohio Supercomputer Center
(OSC). The state-funded entity uses cutting-edge technologies that eventually transpire at a mind-numbing rate of 154 trillion calculations per second.
Imagine what the bionic man would have thought about that.
Since 1987, center researchers have been using better and faster servers to provide supercomputing services and expertise to Ohio industries and universities. The effort supports new research and business opportunities, and ultimately helps boost Ohio's fledgling knowledge economy. Super-speed computing is a competitive force, say its supporters, educating Ohio's workforce in key skills required for future jobs. In addition, companies harnessing the power of the network can simulate real-world product testing quickly and accurately, with the aim of leapfrogging their wares to the top of the market.
A leg up for the 'missing middle'
The OSC, located on Ohio State University's Columbus campus, offers its lighting-fast modeling and state-of-the-art simulation to areas of industry that usually don't have access to such advanced technology.
This "missing middle" is the focus of the National Digital Engineering and Manufacturing Consortium (NDEMC)
program. The venture uses advanced modeling and simulation to boost the prospects for small- and medium-sized supply-chain businesses, says OSC Executive Director Pankaj Shah.
NDEMC is a five-year public-private partnership funded by the Economic Development Administration (EDA). The OSC is involved along with smaller manufacturers and the larger corporations for which they create products, including Ohio powerhouses such as Procter & Gamble
and Lockheed Martin
Most Ohio-based intermediate suppliers don't have the same access to ultra-speed computing as their bigger brethren, notes Shah. Through NDEMC, these companies are able to harness the water-cooled supercomputer bunkered in Columbus. Called the Oakley Cluster
after famous Ohioan Annie Oakley, the computer is actually a cluster of servers linked by cables and software; which is how it achieves such staggering speed. Those 154 trillion calculations per second, incidentally, are called 154 teraflops.
One practical application for the computing behemoth includes crash testing for new automobiles. Instead of taking time to build numerous, costly prototype vehicles from scratch, for example, components can be emulated with swiftness and precision in a simulated environment, says Shah. This does not remove physical testing completely, he notes, but does well to compliment real-world analysis.
"The steps in between can be saved," he says.
Almost 20 small- to medium-sized businesses are paying to use the supercomputer for this work. The cost is well worth the potential benefits for company growth, believes Thomas Lange, director of research and development at Procter & Gamble.
Lange points to a plastic bottle manufacturer in the P&G supply chain that worked with the center to help reproduce a crush test of new bottle designs. Normally, the company would not have had the scale for supercomputing. Programs like NDEMC break that technological barrier, maintains Lange, a critical aspect of keeping up in an ever-shifting global economy where the "other guy" is always racing to get the best product to market first.
"Modeling and simulation is essential for all good manufacturers," says Lange. "Our survival--and the survival of our suppliers--depends on being digitally capable."
Harnessing the superpowers of the supercomputer
Computer modeling over the last 20-plus years since the center came online has evolved to the point where it's more accurate than real-life tests, says Lange. Testing can cover literally thousands of product issues, from taking a tissue out of a box to dishwashing liquid leaking through a laundry pod.
The state, recognizing that further engagement with advanced computing tools was needed for smaller enterprises, awarded a $3 million Innovation Platform Program (IPP)
grant to the OSC and its product partners to design and deploy easy-to-use “manufacturing apps" that can be used remotely. The IPP is part of Ohio Third Frontier
. Client partners are funding the remainder of the three-year, $6.4 million program.
"Imagine apps that can solve very specific science or engineering questions," says Lange. "Now (small companies) can get access to the same computing and software that large companies have every day."
Analyzing something as ordinary as how liquid flows through an automobile manifold would take a specialist in fluid dynamics a week to set up and another week to run, while also exacting a heavy cost from a supplier, notes the R&D director. By pushing a few buttons in an app that models a pipe manifold, the flow patterns and properties of a liquid can be easily simulated.
"That's why (P&G) wanted to be part of this program," Lange says.
Proliferation of advanced manufacturing simulation apps is part of the center's Blue Collar Computing (BCC)
efforts. Launched in 2004, the project provides resources and training to help companies gain easy and affordable access to the newest computing innovations.
Industrial design and engineering firm AltaSim Technologies
, which has worked with companies in fields such as healthcare, defense and aerospace, is one of the firms writing software for the apps. Principal Jeff Crompton sees a wide range of Ohio industries benefitting from supercomputing tools, with everyday consumer suppliers having a chance to benefit from simulation-driven technology as well.
"These are companies that want to re-invent their technology," says Crompton. "It could be a jet engine, or millions of consumer-based products."
Not having to swap out pricey hardware to upgrade your very own HAL 9000 is another benefit of a centralized supercomputer center. "You may get 18 months to two years before your technology is made obsolete," Crompton says. "Even a small system could cost between $30,000 to $40,000 to replace."
The center's computer cluster costs $3 to $5 million to update every few years, reports director Shah. Enhancing the economic outlooks of Ohio manufacturers makes that money well spent in his eyes.
"It's a wonderful resource we're sitting on here," Shah says. "It's a hidden gem."
Lange of P&G likes to take a wider view. Faster, more accurate testing would lead to better products in the market, which in turn could result in Ohio companies stepping out in front of competitors, clearly creating an advantage for a state that's attempting to transform its economy.
"Simulation helps make our products the best they can be," Lange says. "That's our challenge. We don't want people buying stuff from 5,000 miles away that could be made here."